---
layout: null
title: Continuing with R
excerpt: These slides will be presented at the DSCOE Inferno Lite and are intended to give an idea of how to continue with R after an introductory course.
---
DSCOE Inferno Lite
DSCOE Inferno Lite
Ian Kloo
August 2016
Why continue with R?
Reusable Code
Write it once, use it forever
Share with others to avoid duplicate effort
Auditable Methods
Easy to see what you did (assumptions)
Can easily tweak analyses without re-doing everything
Avoids the “hit by a bus” problem
Huge and Growing Community
Open source breeds collaboration
Any question has probably already been asked or answered
Ongoing development guaranteed
What can/should I do Now?
Advanced Modeling
Statistical Methods
Statistical Methods
Base R very good for traditional stats
Bayesian movement alive in R community
Optimization
Many packages, including optimx
Tough syntax, but great auditability
Machine Learning
caret package!
SVM, Bayesian Methods, Random Forest, etc.
Same syntax for everything
Parallel Computing
Apply functions
Snow, Parallel, and many other packages
Take advantage of full local resources
Use AWS to get even more resources
Visualizations
No More Pie Charts!
ggplot
Publication-ready (with some work)
Easy syntax
Lots of support
ggplot: scientific
ggplot: maps
ggplot: heatmaps
ggplot: anything!
Pre-built htmlwidgets
rCharts, networkD3, plotly, leaflet, and many more!
Interactive visualizations
No need to learn javascript
Somewhat difficult to customize (in some cases)
rCharts
Plotly
Leaflet
Sigma
Roll-Your-Own htmlwidgets
Extremely customizable and interactive
Javascript knowledge required
Can write the JS once, then use R!
Add javascript visualizations to presentation, applications, and papers
Difficult learning curve
forceMap
Get a Project!
R is like running - stop for a few weeks and you lose most of it
Quick-turn work projects may not work, but try to use R if possible
Find something you are interested in and work at home